Integrated Foundations of Sensing, Modeling, and Data Assimilation for Sea Ice Prediction

Department of Defense Multidisciplinary University Research Initiatives program through the Office of Naval Research

17 Oct 2022 - Dylan Green

Using our multiple measurement vector approach with an empirically-derived mask in the magnitude prior, we sample the posterior distribution of a complex-valued signal. Here, using four measurement vectors with noise added in the Fourier domain, we show the magnitude of an example image (top left) recovered in three ways: using just the inverse Fourier transform of the mean of the data (top right), finding the mean of the magnitude when a nonempirical prior is used (bottom left), and finding the mean of the magnitude when the empirical prior is used (bottom right).